feat(cuda): hybrid GPU dispatch - fuse dyn + standalone kernels

[0ax1]

Add a hybrid_dispatch module that integrates subtrees with separate kernel dispatch with dynamic dispatch kernels. Subtrees with unsupported encodings (e.g. Zstd) are executed separately and their device buffers are fed back as LOAD ops in the fused plan.

Note that this implicitly enables filtering via the CUDA CUB filter implementation.

[codspeed-hq]

Merging this PR will degrade performance by 10.2%

❌ 1 regressed benchmark
✅ 1015 untouched benchmarks
⏩ 1522 skipped benchmarks¹

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	map_each[BufferMut<i32>, 128]	770.6 ns	858.1 ns	-10.2%

---

_{Comparing ad/fallback (c5a1ed5) with develop (1c8667c)}

Footnotes

1. 1522 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩